Acute myeloid leukemia (AML) is one of the most common forms of cancer in children and adults. CD33, a protein often expressed on the surface of leukemic cells, has been a long-standing target for AML treatment. Gemtuzumab ozogamicin (GO), known commercially as Mylotarg, is an antibody-fused drug conjugate targeting CD33. However, not all patients with CD33+ AML respond to the drug for unknown reasons. One possibility is that they express a splice variant of CD33 that allows escape from the antibody. One splice variant of CD33 is associated with a single nucleotide polymorphism, where the minor "T" allele mediates higher translation of CD33 without exon 2 (CD33Δ2) than the dominant "C" allele. Previous studies have shown that addition of GO treatment to standard of care was only effective in AML patients with two "C" alleles. Around 50% of population will have at least one "T" allele, leading to many patients that potentially do not respond well to GO treatment. Therefore, a large focus has been developing treatment options for these patients, including antibody-based treatments. The Walter lab (Clinical Research Division) was interested in exploring antibody-based designs for CD33Δ2 – their results were recently published in Leukemia.
To start their research, the authors had to make antibodies to CD33Δ2, as commercially available ones did not exist. They used hybridoma techniques to screen and purify antibodies. Using a control cell lines that were CD33Δ2 negative or engineered to express either full length CD33 (CD33FL) or CD33Δ2, they found two antibodies that bound CD33Δ2 but not CD33FL. Next, they sought to use their antibodies to determine if AML lines derived from patients with the "C" or "T" alleles expressed CD33Δ2. However, they did not see any expression of CD33Δ2 on the cell surface of lines with either genotype by flow cytometry. Additionally, they screened primary AML samples from patients with each genotype. Again, they did not see expression of CD33Δ2 on the cell surface, though patients with either allele showed expression of CD33FL. The only time cell surface expression of CD33Δ2 was demonstrated was with cell lines engineered to overexpress CD33Δ2 from a viral vector.
The signal sequence required for CD33 transit to the cell surface lies mostly in exon 1, however four amino acids are in exon 2. Therefore, it was possible that CD33Δ2 did not make it to the cell surface due to lack of a functional signal sequence. In order to test this idea, the authors permeabilized their cells prior to antibody use, a technique that allows for detection of surface and intracellular proteins. They confirmed that permeabilization did not disrupt antibody binding to CD33Δ2, as cells that overexpressed the protein still showed strong signal. However, permeabilized AML cell lines and primary samples did not show expression of CD33Δ2. Finally, in order to confirm that CD33Δ2 protein was completely absent, the authors used an antibody to the C-terminus of CD33 to capture all CD33FL and CD33Δ2 from cell lysates. Western blot analysis of the CD33-captured cell lysate showed that only CD33FL was expressed in AML cell lines and primary samples. CD33Δ2 signal was only present in cells engineered to overexpress the protein. These data suggest that while the CD33Δ2 RNA transcripts exist in AML cells, they fail to make CD33Δ2 protein that could be targeted with antibody-based therapy. While this research largely represents co-called "negative data," it is a careful and through evaluation of a target with wide interest. Rigorous elimination of potential therapeutic targets is valuable, as it allows researchers to spend more time evaluating other potentially effective targets.
This work was supported by the Leukemia & Lymphoma Society, the National Institutes of Health, the American Society of Clinical Oncology/Conquer Cancer Foundation, and Alex’s Lemonade Stand Foundation.
Fred Hutch/UW Cancer Consortium members Brent Wood, Hans-Peter Kiem, and Roland Walter contributed to this work.
Godwin CD, Laszlo GS, Wood BL, Correnti CE, Bates OM, Garling EE, Mao ZJ, Beddoe ME, Lunn MC, Humbert O, Kiem HP, Walter RB. 2020. The CD33 splice isoform lacking exon 2 as therapeutic target in human acute myeloid leukemia. Leukemia. doi: 10.1038/s41375-020-0755-7